[LAST CALL] General Assembly Authorship Guidelines

[Daan Island]

As part of their six point GA overhaul plan, mods nuked the old ruleset and built a new one from the ground up. Many people made salient points about issues with this ruleset, which were entirely ignored. Fast forward a few months and they're rebuilding the Relevance rule from first principles after the blank repeal. As such, I bring to you a declaration concerning a community standard for General Assembly authors. The aim is to have this be enough of a community consensus to have it effectively apply by default, and I believe that is mostly already the case. Nonetheless, I think it would be optimal to formalize this as a declaration. Of course, this is not intended to be something people get crucified over if they have one similar clause to a previous resolution, just a bare minimum that should be trivial to follow if you do your due diligence. One big advantage here is formalizing blockers though clause 1 of the main list.

Someone will inevitably bring up the issue of legality due to rule 3a, but I think this is legal as written. It doesn't explicitly mention the rules overhaul or even frame itself specifically as a fix to that, it exists independently to any site staff decision. In July last year Sedge wrote:

A ruleset along the lines of the one in Violet's proposal would be much easier for players to comply with, and it leaves decisions on standards directly in the hands of players.

General Assembly Authorship Guidelines

category: Declaration | Co-authors: Cretox State

---

The Security Council,

Valuing certain practices in General Assembly proposal drafting as fostering a healthier, more democratic, and more efficient international legislative body,

Finding that formal and informal rules have both varied greatly over the existence of the General Assembly and its predecessor which shall not be named, often being driven by international community standards,

Wishing therefore to adopt a formal community standard more stringent than current legal rules and timeless in its domain,

Defines such a community standard as follows, which nations are urged to conform to when writing General Assembly legislation:

1. Proposals and their mandates should not be contradictory to those of extant resolutions; contradicting laws significantly undermine the integrity and enforceability of the whole body of General Assembly law,
2. Proposals should not overtly mention or advertise any nation or region; the Security Council already provides mechanisms to recognize deserving nations and regions while the General Assembly risks losing legislative credibility if it is hijacked for such purposes,
3. Proposals should not duplicate the contents or mandates of standing resolutions; requiring the same acts or omissions multiple times produces entirely unnecessary complications in logistics, enforcement, and potentially other areas,
4. Repeals of passed resolutions should make a good faith effort not to misrepresent the target resolution; arguing for the repeal of standing legislation should be feasible using truthful information if it is in fact flawed or unnecessary,
5. Repeals should contain an argument for repeal which is relevant to the specific contents of the target resolution; such helps filter away vague assertions which could apply to any resolution,

Promotes the following additional guidelines for authoring nations to follow, as a matter of common courtesy:

1. Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,
2. No proposal should be written with the aid of artificial intelligence; the Security Council has previously denounced the use of artificial intelligence for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

As much as I support you and ethnon, I think that this just highlights already established guidelines.. so sorry, but, AGAINST

[Heavens Reach]

As much as I support you and ethnon, I think that this just highlights already established guidelines.. so sorry, but, AGAINST

No guidelines have been established for how the General Assembly operates post-reset or under the new rules, particularly not as a statement of community expectations.

[Grenartia]

The discussion is turning out not to be so simple, and two competing visions for tweaking the rule are being considered.

Out of curiosity, has there been any movement on this?

And what are the two competing visions?

[Lipizia]

Ok, For.
Anyway, I suggest to use "World Assembly" instead of "General Assembly": if there a SC proposal have been submitted, and it's completely made with IA, it's ok because it's SC and not GA?

[Saint Halo]

Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,

What would you consider "lengthy?" Does that amount of time changer per author?

[Northern Slatavia]

Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,

What would you consider "lengthy?" Does that amount of time changer per author?

Well GA#6 took almost 3 months of drafting, and GA#7 took a little less than two weeks. Both recieved a recommendation for the vote from the author of this proposal. I think it's really matters what the topic is, as a resolution on War Crimes will most likely take more time than one on limiting power.

[Islandwalk]

I have implemented Grenartia's feedback and slightly modified the phrasing as follows (relevant section underlined):

No proposal should be written with the aid of artificial intelligence; the Security Council has previously denounced the use of artificial intelligence for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

No proposal should be written using large language models; the Security Council has previously denounced the use of artificial intelligence for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

---

Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,

What would you consider "lengthy?" Does that amount of time changer per author?

As this is a matter of common courtesy and not a firm rule, I avoid specifying a strict standard amount. The real answer is that it depends on the topic. Something simple like a liberation repeal (the non-contentious ones) can be up for 2 weeks before submission, first time GA proposals and the like would benefit from at least a month of drafting, and more if feedback is scarce.

[Dengist China]

I have implemented Grenartia's feedback and slightly modified the phrasing as follows (relevant section underlined):

No proposal should be written with the aid of artificial intelligence; the Security Council has previously denounced the use of artificial intelligence for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

No proposal should be written using large language models; the Security Council has previously denounced the use of artificial intelligence for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

I think it should be the following instead, which is editing Grenartia's edit:

No proposal should be written using large language models (LLMs); the Security Council has previously denounced the use of LLMs for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

[Heavens Reach]

I have implemented Grenartia's feedback and slightly modified the phrasing as follows (relevant section underlined):

I think it should be the following instead, which is editing Grenartia's edit:

No proposal should be written using large language models (LLMs); the Security Council has previously denounced the use of LLMs for creative and diplomatic purposes in SC#578 for its theft of ideas and use in achieving recognition without authentic effort.

I may as well follow up on this. Specifying LLMs is a mistake. Generative AI (GAI) is the more general use case and encompasses LLMs and future generative architectures.

[Dengist China]

I may as well follow up on this. Specifying LLMs is a mistake. Generative AI (GAI) is the more general use case and encompasses LLMs and future generative architectures.

Fair point. I agree.

[Adriatic Democratic Empire]

my main questions regarding this have been answered so I'm pretty comfortable voting FOR

[Nolrem]

my main questions regarding this have been answered so I'm pretty comfortable voting FOR

If they ever let us vote on it

[Hesitant Dispatch 440]

my main questions regarding this have been answered so I'm pretty comfortable voting FOR

If they ever let us vote on it

Jury's still out on that.

[Grenartia]

Thank you for the nod, Islandwalk.

I think it should be the following instead, which is editing Grenartia's edit:

I may as well follow up on this. Specifying LLMs is a mistake. Generative AI (GAI) is the more general use case and encompasses LLMs and future generative architectures.

The primary reason I avoid "GAI" or even "GenAI" for "Generative AI" is because it shares the same abbreviation as "General AI", which is widely understood to refer to the kind of AI that many people would understand to be 'deserving of rights as a person'.

That's also why I specified Large Language Models. To be clear about the kind of technology that we're trying to condemn the use of. While all LLMs are 'generative AI', not all 'generative AI' are LLMs, but only LLMs are capable of being used to generate textslop. AFAIK, other forms of generative AI generate imageslop, which isn't relevant for stopping slop proposals. I'm not opposed to another resolution to condemn imageslop on here, but I think it should be a separate resolution.

If they ever let us vote on it

Jury's still out on that.

How do we poke Sedge to get an update, anyways?

[Dushina]

Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,

Should this clause make explicit exception for cases of especially time-sensitive proposals, or is this something wholly unacceptable in the GA? If so, why or why not? I do admittedly share New Halo's concern in this regard.

[Heavens Reach]

Thank you for the nod, Islandwalk.

I may as well follow up on this. Specifying LLMs is a mistake. Generative AI (GAI) is the more general use case and encompasses LLMs and future generative architectures.

The primary reason I avoid "GAI" or even "GenAI" for "Generative AI" is because it shares the same abbreviation as "General AI", which is widely understood to refer to the kind of AI that many people would understand to be 'deserving of rights as a person'.

That's also why I specified Large Language Models. To be clear about the kind of technology that we're trying to condemn the use of. While all LLMs are 'generative AI', not all 'generative AI' are LLMs, but only LLMs are capable of being used to generate textslop. AFAIK, other forms of generative AI generate imageslop, which isn't relevant for stopping slop proposals. I'm not opposed to another resolution to condemn imageslop on here, but I think it should be a separate resolution.

Jury's still out on that.

How do we poke Sedge to get an update, anyways?

The term you are concerned about is "artificial general intelligence" (AGI). That is a distinct acronym, and has a distinct meaning from "generative artificial intelligence" (GAI).

Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,

Should this clause make explicit exception for cases of especially time-sensitive proposals, or is this something wholly unacceptable in the GA? If so, why or why not? I do admittedly share New Halo's concern in this regard.

I cannot think of a scenario in which any GA proposal has ever been treated with that sort of urgency or under which it should be. Almost invariably, except for when it's in service of a power play, submitting straight to the queue renders your proposal DOA.

[Grenartia]

Thank you for the nod, Islandwalk.



The primary reason I avoid "GAI" or even "GenAI" for "Generative AI" is because it shares the same abbreviation as "General AI", which is widely understood to refer to the kind of AI that many people would understand to be 'deserving of rights as a person'.

That's also why I specified Large Language Models. To be clear about the kind of technology that we're trying to condemn the use of. While all LLMs are 'generative AI', not all 'generative AI' are LLMs, but only LLMs are capable of being used to generate textslop. AFAIK, other forms of generative AI generate imageslop, which isn't relevant for stopping slop proposals. I'm not opposed to another resolution to condemn imageslop on here, but I think it should be a separate resolution.



How do we poke Sedge to get an update, anyways?

The term you are concerned about is "artificial general intelligence" (AGI). That is a distinct acronym, and has a distinct meaning from "generative artificial intelligence" (GAI).

I'll concede on the AGI thing, but I'm still unconvinced that specifying LLMs is a bad move.

Proposals should be drafted in public forums for a lengthy enough time period to receive feedback from other nations; public drafting and consideration of feedback results in less instances of faulty writing and gives WA members a chance to openly voice their concerns,

Should this clause make explicit exception for cases of especially time-sensitive proposals, or is this something wholly unacceptable in the GA? If so, why or why not? I do admittedly share New Halo's concern in this regard.

I cannot think of a scenario in which any GA proposal has ever been treated with that sort of urgency or under which it should be.

I concur.